Red Hat Bugzilla – Bug 1270834
Remove network disk or direct lun times out
Last modified: 2016-02-10 14:21:48 EST
Created attachment 1082020 [details]
Description of problem:
HotUnplug of a cinder disk fails
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Have a VM+cinder disk
3.attempt to deactivate the disk while VM is up
Target release should be placed once a package build is known to fix a issue. Since this bug is not modified, the target version has been reset. Please use target milestone to plan a fix for a oVirt release.
A couple of questions to investigate the issue:
* With which disk interface has the Cinder disk been attached?
* Which OS was installed?
* Which libvirt/qemu versions?
* Can you please attach the relevant VDSM log as well
After discussing with Amit, it seems that the hot-unplug timeout indication is based on the serial xml element: https://gerrit.ovirt.org/#/c/45138/19/vdsm/virt/vm.py (line 2663). However, 'serial' element is missing on Cinder and LUN disks.
@Nir - how do you suggest to tackle this issue? Should we have a unique ID be each disk or perhaps have some another technique to identify unplugged disk.
(In reply to Daniel Erez from comment #3)
> After discussing with Amit, it seems that the hot-unplug timeout indication
> is based on the serial xml element:
> https://gerrit.ovirt.org/#/c/45138/19/vdsm/virt/vm.py (line 2663). However,
> 'serial' element is missing on Cinder and LUN disks.
> @Nir - how do you suggest to tackle this issue? Should we have a unique ID
> be each disk
Yes, why don't we set the serial element?
We should use the same way for any kind of disk.
It seems this bug is not related directly to Cinder implementation since it is reproducing with Dlun disks, this case was passing storage automation and now it fails, thus marking it, Regression :))
This bug report has Keywords: Regression or TestBlocker.
Since no regressions or test blockers are allowed between releases, it is also being identified as a blocker for this release. Please resolve ASAP.
removing blocker of https://bugzilla.redhat.com/1270834, since this appears to be a general issue
yes it is not blocking bz #1135132, sorry for yelling
First, the fact that direct lun does not have a serial does not explain
the hot unplug failure - I would expect that the operation will always succeed, because when we check if the disk exists, we will never find it
(there is no disk with empty serial).
Second, you cannot detach the disk the vm was booted from - this will not
work for any type of disk, and is not a bug.
It seems that the description for this bug is incorrect, and the logs are
Ori, please attach vdsm logs showing the failure, as requested in comment 2.
Correction, qemu does allow hot-unplugging the boot disk. After this, it
is not possible to connect to the vm console, it is busy looping,
complaining about reeadonly file system.
This test should be performed on a second disk, not on the boot disk.
We should check with qemu/libvirt guys about allowing hot-unplugging of the boot disk, this does not look useful to any user.
In oVirt testing is done on single release by default. Therefore I'm removing the 3.6 flag. If you think this bug must be tested in 3.6 as well, please re-add the flag. Please note we might not have testing resources to handle the 3.6 clone.
Bug tickets that are moved to testing must have target release set to make sure tester knows what to test. Please set the correct target release before moving to ON_QA.
Verified using vt18.2 (vdsm-4.16.29-1.el7ev.x86_64)
1. Have a VM and OS installed (Boot disk on Gluster domain)
2. Hotplug direct lun (over ISCSI)
3. HotUnPlug the direct lun
oVirt 3.5.6 has been released and the bz verified, moving to closed current release.